Abstract
This paper describes a practical method for estimating the contact condition between wheel and rail using an instrumented wheelset in a railway vehicle. In a previous study, it was confirmed that an observer based on Kalman filter can estimate wheelset motion variables from measurement data of the instrumented wheelset using a single-wheel creep-force model. However, in the previous method, there is still a problem in that a precise value of friction coefficient between wheel and rail is required as a model parameter to estimate the wheelset motion variables with high accuracy. The accuracy of the value of friction coefficient affects the accuracy of estimation values since a single-wheel creep-force model refers to the creep force and the creepage. Moreover, the decision of the creep force heavily depends on the friction coefficient. To solve the problem, this study proposes a method to estimate the wheelset motion variables and the friction coefficient simultaneously by modifying the plant model. The estimation performance of the proposed method is verified by vehicle dynamics simulations and running test data. The verification results show that the proposed method is effective in estimating the friction coefficient when the track curve radius is smaller than or equal to 600m. In addition, the proposed method can improve the estimation accuracy of the wheelset motion variables compared to the previous method.
